On some combinatorial problems in metric spaces of bounded doubling dimension

Save this PDF as:
 WORD  PNG  TXT  JPG

Size: px
Start display at page:

Download "On some combinatorial problems in metric spaces of bounded doubling dimension"

Transcription

1 On some combinatorial problems in metric spaces of bounded doubling dimension Michiel Smid February 5, 2010 Abstract A metric space has doubling dimension d if for every ρ > 0, every ball of radius ρ can be covered by at most 2 d balls of radius ρ/2. This generalizes the Euclidean dimension, because the doubling dimension of Euclidean space R d is proportional to d. The following results are shown, for any d 1 and any metric space of size n and doubling dimension d: First, the maximum number of diametral pairs is Θ(n 2 ). Second, if d = 1, the maximum possible weights of the minimum spanning tree and the all-nearest neighbors graph are Θ(R log n) and Θ(R), respectively, where R is the minimum radius of any ball containing all elements of the metric space. Finally, if d > 1, the maximum possible weights of both the minimum spanning tree and the all-nearest neighbors graph are Θ(Rn 1 1/d ). These results show that, for 1 d 3, metric spaces of doubling dimension d behave differently than their Euclidean counterparts. 1 Introduction Let S be a set of n points in R d, where d 1 is a constant. Let G be a graph with vertex set S, in which the weight of any edge (p, q) is defined to be the Euclidean distance pq between p and q. The weight wt(g) of G is defined to be the sum of the weights of the edges in G. School of Computer Science, Carleton University, Ottawa, Canada. This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). 1

2 We are interested in the maximum values (over all sets of n points) of the number of diametral pairs, the weight of the minimum spanning tree, and the weight of the all-nearest neighbors graph. Diametral pairs: The diameter of S is defined to be the maximum distance between any pair of points in S. Two points p and q in S form a diametral pair, if pq is equal to the diameter of S. Let DP(S) be the number of diametral pairs in the set S. Obviously, DP(S) = 1 if d = 1. If d = 2, the maximum possible value of DP(S) is n, whereas if d = 3, the maximum possible value is 2n 2. If d 4, the value of DP(S) can be as large as Θ(n 2 ). For proofs of these claims, see Chapter 13 in Pach and Agarwal [5]. Minimum spanning tree: If S is contained in the d-dimensional cube [0, L] d, then the weight wt(mst (S)) of its minimum spanning tree MST (S) is O(n 1 1/d L). This result was first shown by Few [2]. Subsequently, Smith [7], and Steele and Snyder [8] gave tighter analyses in terms of the constant factor in the Big-Oh bound. For the set S consisting of the n vertices on an n 1/d... n 1/d grid with cells having sides of length n 1/d L, we have wt(mst (S)) = Ω(n 1 1/d L). Thus, the above upper bound on wt(mst (S)) is tight. All-nearest neighbors graph: For any point p in S, let NN S (p) denote a nearest neighbor of p in S. That is, NN S (p) is a point q in S \ {p} for which the Euclidean distance pq is minimum. The all-nearest neighbors graph ANN (S) is defined to be the directed graph with vertex set S and edge set {(p, NN S (p)) : p S}. Assume again that the point set S is contained in the cube [0, L] d. Since ANN (S) is a subgraph of MST (S), we have wt(ann (S)) = O(n 1 1/d L). The grid example given above shows that this upper bound is tight. All results mentioned above are valid in the d-dimensional Euclidean space. A natural question to ask is whether or not similar results hold in an arbitrary metric space. Let S be a finite set of elements, called points, that has a distance function associated with it. This function assigns to any two points p and q in S a real number pq, which is called the distance between p and q. The set S is called a metric space, if for all points p, q, and r in S, 2

3 1. pp = 0, 2. if p q, then pq > 0, 3. pq = qp, and 4. pq pr + rq. The fourth property is called the triangle inequality. If c is a point in S and R > 0 is a real number, then the ball with center c and radius R is defined to be the set {p S : cp R}. The radius of a subset S of S is defined to be the minimum radius of any ball that contains S. Observe that the center of this ball is not necessarily an element of S. For example, the set S of n points, in which each pair of distinct points has distance R, is a metric space of radius R. Obviously, we have DP(S) = ( n 2), wt(mst (S)) = (n 1)R, and wt(ann (S)) = nr. Moreover, these are the largest possible values in terms of n and R. Thus, in order to get non-trivial results, we have to consider restricted classes of metric spaces. In this paper, we consider metric spaces of bounded doubling dimension, which were introduced by Assouad [1]; see also the book by Heinonen [4]. Definition 1 Let S be a finite metric space and let λ be the smallest integer such that the following is true: For each real number ρ > 0, every ball in S of radius ρ can be covered by at most λ balls of radius ρ/2. The doubling dimension of the metric space S is defined to be log λ. Observe that every metric space of n points has doubling dimension at most log n. Furthermore, it is not difficult to prove that the doubling dimension of the Euclidean metric in R d is proportional to d. In this paper, we prove the following results, for any metric space S consisting of n points and having doubling dimension d: 1. If d 1, then DP(S) is Θ(n 2 ) in the worst case. Thus, the maximum possible value of DP(S) is very different from the corresponding value in Euclidean space R d for 1 d 3. In fact, we show that, for d = 1, the maximum possible value of DP(S) is n 2 /4. For d 4, the bound coincides with the corresponding bound for R d. 3

4 2. If d = 1, then wt(mst (S)) is Θ(R log n) in the worst case, whereas wt(ann (S)) is Θ(R) in the worst case. Thus, metric spaces of doubling dimension 1 behave differently than their one-dimensional Euclidean counterparts. 3. If d > 1, then both wt(mst (S)) and wt(ann (S)) are Θ(Rn 1 1/d ) in the worst case. Thus, these bounds coincide with the corresponding bounds for Euclidean space R d. Related work: In recent years, metric spaces of bounded doubling dimension have received a lot of attention in the algorithms literature; see Har-Peled and Mendel [3] and the references given in that paper. In [3], it was already observed that these spaces are different from their Euclidean counterparts: In Euclidean space R d, the all-nearest neighbors graph can be computed in O(n log n) time, whereas the worst-case running time becomes Θ(n 2 ) even if the doubling dimension is equal to 1. Smid [6] defined the following metric space, based on an earlier example in [3]: Let S = {p 1, p 2,..., p n } and let 0 < ɛ < 1 be a real number. For each i and j with 1 i j n, we define 0 if i = j, p i p j = p j p i = 4 j if i = 1 and j > 1, 4 j + ɛ otherwise. Then S is a metric space of doubling dimension 1 and radius R = 4 n + ɛ. We have NN S (p 1 ) = p 2 and, for each j with 2 j n, NN S (p j ) = p 1. Thus, both ANN (S) and MST (S) are equal to the star-graph consisting of all edges (p 1, p j ), 2 j n. (In contrast, it is well known that in Euclidean space R d, the maximum degrees of both these graphs are upper bounded by a function that depends only on d.) Both wt(ann (S)) and wt(mst (S)) are close to 4R/3. The rest of this paper is organized as follows. In Section 2, we prove the upper bounds on DP(S), wt(ann (S)), and wt(mst (S)), for any finite metric space of doubling dimension d. The proofs will be by induction on the number of points in S. Some care has to be taken, because the doubling dimension d of a subset of S may be larger than d. We show that d is always at most 2d. 4

5 In Section 3, we prove that the upper bounds in Section 2 are tight. The lower bounds for DP(S), wt(ann (S)), and wt(mst (S)) are in fact obtained by one single class of metric spaces. We conclude in Section 4 by presenting some more results that show that metric spaces of doubling dimension 1 are very different than the Euclidean space R 1. 2 The upper bounds We start by showing that DP(S) n 2 /4 for every metric space S of size n and doubling dimension 1. The proof uses the following lemma which states that in any isosceles triangle, the base is shorter than the two sides of equal length. Lemma 1 Let S be a finite metric space of doubling dimension 1, and let p, q, and r be three pairwise distinct points in S such that pq = pr. Then qr < pq. Proof. The proof is by contradiction. Let p, q, and r be a smallest triple of points for which the lemma does not hold. That is, let l be the smallest real number such that there exist three points p, q, and r in S such that l = pq = pr and qr l. Let B be the ball with center p and radius l. Observe that B contains the points p, q, and r. Since S has doubling dimension 1, we can cover B by two balls B 1 and B 2 centered at c 1 and c 2, respectively, and having radius l/2. We may assume without loss of generality that p is contained in B 1. Let l = qr. If both q and r are contained in B 2, then l = qr qc 2 + c 2 r l/2 + l/2 = l l. It follows that c 2 q = c 2 r = l/2. Since qr = l l/2, the triple c 2, q, and r forms a smaller counterexample to the lemma, which is a contradiction. Thus, q and r are not both contained in B 2. We may assume without loss of generality that q is contained in B 1. We have l = pq pc 1 + c 1 q l/2 + l/2 = l, which implies that c 1 p = c 1 q = l/2. Since pq = l l/2, the triple c 1, p, and q forms a smaller counterexample to the lemma, which is again a contradiction. 5

6 Theorem 1 Let S be a metric space of doubling dimension 1 and consisting of n points. Then DP(S) n 2 /4. Proof. Let R be the diameter of S, let c be a point in S, and let B be the ball with center c and radius R. Observe that B contains all points of S. Cover B by two balls B 1 and B 2, both having radius R/2, and define S 1 = S B 1 and S 2 = S \ S 1. Assume that S 1 contains two points p and q such that pq = R. Then, denoting the center of B 1 by c, R = pq pc + cq R/2 + R/2 = R and, therefore, cp = cq = R/2, which contradicts Lemma 1. Thus, the diameter of S 1 is less than R. By the same argument, the diameter of S 2 is less than R. If we denote the size of S 1 by m, then it follows that DP(S) is at most m(n m). Since the function m(n m), for 0 m n, is maximized when m = n/2, the proof is complete. The proofs of the upper bounds on the weights of MST (S) and ANN (S) will be by induction on the size of the metric space S. We have to be careful, however, because the doubling dimension d of a subset S of S may be larger than the doubling d dimension of S. The reason is that the value of d is determined by only considering balls that are centered at points of S. To give an example, let n = m 2 and let S be a metric space of size n, in which the distance between any two distinct points is equal to 2. The doubling dimension d of S is equal to d = log n = 2 log m. Partition S into subsets S 1, S 2,..., S m, each consisting of m points. Define S = S {p 1, p 2,..., p m }, where the p i s are new points. For each i with 1 i m, and for each q in S i, we define p i q = qp i = 1. For any other pair of distinct points in S, their distance is defined to be 2. Observe that S is a metric space. We claim that the doubling dimension d of S is equal to d = log(m + 1). To prove this claim, let ρ > 0 be a real number and let B be a ball in S of radius ρ. If ρ 2, then B is covered by m balls of radius ρ/2, centered at the points p 1, p 2,..., p m. If 1 ρ < 2, then B contains at most m + 1 points and, therefore, we can cover B by m + 1 balls of radius ρ/2. Finally, if ρ < 1, then B is a singleton set and we can cover B by one ball of radius ρ/2. Thus, d is indeed equal to log(m + 1). 6

7 For large values of m, the ratio d /d converges (from below) to 2. The next lemma states that this ratio can never be larger than 2. Lemma 2 Let S be a finite metric space of doubling dimension d and let S be a non-empty subset of S. Then the doubling dimension of the metric space S is at most 2d. Proof. Let c be a point of S, let ρ > 0 be a real number, and let B = {p S : cp ρ}. Thus, B is a ball in the metric space S. Consider the corresponding ball B in S, i.e., B = {p S : cp ρ}. Since S has doubling dimension d, we can cover B by balls B 1, B 2,..., B k, all having radius ρ/4, where k 2 2d. For 1 i k, let c i be the center of B i. If c i S, then we define B i to be the ball in S of radius ρ/2 that is centered at c i. If c i S and B i S, then we define B i to be the ball in S of radius ρ/2 that is centered at an arbitrary point of B i S. We claim that the collection of balls {B i : 1 i k, B i S } cover the ball B. To prove this, let p be a point in B. Then p B and, thus, there is an index i such that p B i. If c i S, then p B i. Otherwise, let c i be the center of B i. Since c ip c ic i + c i p ρ/4 + ρ/4 = ρ/2, it follows that p B i. Thus, we have covered the ball B in S by at most 2 2d balls (again, in S ) of radius ρ/2. Theorem 2 Let S be a metric space consisting of n points, let d be the doubling dimension of S, and let R be the radius of S. 1. If d = 1, then wt(mst (S)) 2R log n. 2. If d > 1, then wt(mst (S)) 12Rn 1 1/d. Proof. We start by considering the case when d = 1. We will prove the following claim by induction on the size of V : Let V be a non-empty subset of S and let R V be the radius of V. Then wt(mst (V )) 2R V log V. (1) By taking V = S, this claim will prove the first part of the theorem. 7

8 If V is a singleton set, then (1) obviously holds. Assume that V contains at least two elements and, further, assume that (1) holds for all non-empty subsets of S having less than V elements. Let B be a ball of radius R V that contains V. Since the doubling dimension of S is equal to one, we can cover B by two balls B 1 and B 2, both having radius R V /2. Note that the centers of B 1 and B 2 are not necessarily points of V. Define V 1 = V B 1 and V 2 = V \ V 1. Then both V 1 and V 2 are non-empty (otherwise, the radius of V would be less than R V ). Let T be the spanning tree of V consisting of MST (V 1 ), MST (V 2 ), and one edge joining an arbitrary point of V 1 with an arbitrary point of V 2. Let R V1 and R V2 be the radius of V 1 and V 2, respectively. Then R V1 R V /2 and R V2 R V /2. Since the diameter of V is at most 2R V, and by applying the induction hypothesis, we obtain wt(t ) 2R V1 log V 1 + 2R V2 log V 2 + 2R V R V (log V 1 + log V 2 + 2). Since the function log x is convex for x > 0, and since V 1 + V 2 = V, it follows that wt(t ) R V (log( V /2) + log( V /2) + 2) = 2R V log V. Since wt(mst (V )) wt(t ), we have shown that (1) holds for V. In the rest of the proof, we consider the case when d > 1. We will prove the following claim, again by induction on the size of V : Let V be a non-empty subset of S and let R V be the radius of V. Then wt(mst (V )) 12R V V 1 1/d 12R V. (2) By taking V = S, this claim will prove the second part of the theorem. If V is a singleton set, then wt(mst (V )) = R V = 0 and (2) holds. Assume that V consists of two points p and q. Then wt(mst (V )) = pq 2R V. Thus, (2) holds if we can show that /d 12. This inequality can be rewritten as 7 2 1/d 12. The latter inequality holds, because d log 3. From now on, we assume that V contains at least three elements and (2) holds for all non-empty subsets of S having less than V elements. Let B be a ball of radius R V that contains V. Since the doubling dimension of S is equal to d, we can cover B V by balls B 1, B 2,..., B k, all having radius R V /2, where k 2 d. Observe that, since the radius of V is equal to R V, we 8

9 have k 2. For 1 i k, let V i be the set of points of V that are in the ball B i. In case a point of V is contained in more than one ball, we put it in exactly one subset V i. Thus, the sets V 1, V 2,..., V k are pairwise disjoint and their union is equal to V. We may also assume that none of the sets V i is empty. We next claim that we may assume that k 3. Indeed, if k = 2, then, since V 3, one of the sets V 1 and V 2, say V 2, contains at least two points. We partition V 2 into two disjoint non-empty subsets, which we call V 2 and V 3, and, thus, we can set k = 3. Let i be an index with 1 i k, and let R Vi be the radius of the set V i. Then R Vi R V /2. By the induction hypothesis, we have Since V i 1 1/d 1 0, it follows that wt(mst (V i )) 12R Vi ( Vi 1 1/d 1 ). wt(mst (V i )) 6R V ( Vi 1 1/d 1 ). For each i with 1 i k, choose an arbitrary point in V i, and let T be an arbitrary spanning tree of these k points. Since the diameter of V is at most 2R V, we have wt(t ) 2(k 1)R V 2kR V. Let T be the spanning tree of V consisting of T and MST (V 1 ), MST (V 2 ),..., MST (V k ). Then wt(t ) k ( 6R V Vi 1 1/d 1 ) + 2kR V i=1 k = 6R V V i 1 1/d 4kR V i=1 k 6R V V i 1 1/d 12R V. i=1 Since the function x 1 1/d is convex for x > 0, and since k i=1 V i = V, it follows that k V i 1 1/d i=1 k ( V /k) 1 1/d = k 1/d V 1 1/d 2 V 1 1/d. i=1 9

10 Thus, we have wt(t ) 12R V V 1 1/d 12R V. Since wt(mst (V )) wt(t ), we have shown that (2) holds for V. Theorem 3 Let S be a metric space consisting of n points, where n 2, let d be the doubling dimension of S, and let R be the radius of S. 1. If d = 1, then wt(ann (S)) 4R. 2. If d > 1, then wt(ann (S)) 12Rn 1 1/d. Proof. Since ANN (S) is a subgraph of MST (S), the second claim follows from Theorem 2. Assume that d = 1. The first claim will follow from the following more general claim: For any subset V of S with V 2 and having radius R V, p, NN S (p) 4R V. (3) p V The proof is by induction on the size of V. If V consists of two points a and b, then p V p, NN S(p) 2 ab 4R V, thus (3) holds. Assume that V contains at least three elements and, further, assume that (3) holds for all subsets of S having at least two and less than V elements. Let B be a ball of radius R V that contains V. We cover B by two balls B 1 and B 2, both having radius R V /2, and define V 1 = V B 1 and V 2 = V \ V 1. Observe that both V 1 and V 2 are non-empty. Let R V1 and R V2 be the radius of V 1 and V 2, respectively. Then R V1 R V /2 and R V2 R V /2. If both V 1 and V 2 contain at least two points, then, using the induction hypothesis, p, NN S (p) = p, NN S (p) + p, NN S (p) p V 1 p V 2 p V 4R V1 + 4R V2 4R V. Assume that one of V 1 and V 2, say V 1, consists of one point a. Since the set V has radius R V, we have a, NN S (a) R V. Combining this with the induction hypothesis, we obtain p, NN S (p) = a, NN S (a) + p, NN S (p) R V + 4R V2 4R V. p V 2 p V 10

11 This completes the proof. 3 The lower bounds In this section, we show that the upper bounds in Theorems 1, 2 and 3 are tight. In fact, we give an example of one class of metric spaces that achieve the upper bounds in all these theorems. This class is obtained by repeatedly applying the following transformation: Let m 2 be an integer and assume we are given a metric space S consisting of m points. Let d be the doubling dimension of S, let R be the radius of S, and assume that the diameter of S is equal to R. We define a new metric space S in the following way. For each i with 1 i 2 d, let S i be a copy of the metric space S, and let S be the union of these copies. For any two points p and q in S, if they are in the same copy S i, then the distance pq (in S) is defined to be their distance in S i. Otherwise, we define pq to be 2R. Let n be the size of S, so that n = 2 d m. The following claims can easily be verified. The triangle inequality holds for the set S and, thus, S is a metric space. Both the diameter and the radius of S are equal to R = 2R and ( 2 d DP(S) = 2 ) m 2 = 1 2 (1 12 d ) n 2. (4) Since the diameter of each set S i is equal to R, and since distances between points in S i and points in S j (for j i) are equal to 2R, we have wt(ann (S)) = 2 d wt(ann (S )). (5) By running Kruskal s minimum spanning tree algorithm on S, it follows that wt(mst (S)) = (2 d 1)R + 2 d wt(mst (S )). (6) We finally claim that the doubling dimension of S is equal to d. To prove this, let c be a point in S, let ρ > 0 be a real number, and consider the ball B = {p S : cp ρ}. First assume that ρ < 2R. Then B is completely contained in one of the sets S i. Since the doubling dimension of S i is equal to d, and since distances 11

12 in S i are the same as distances in S, it follows that B can be covered by at most 2 d balls of radius ρ/2. Now assume that ρ 2R. Since the radius of S i is equal to R, there exists a ball B i of radius R that contains S i. The union of these balls B i, for 1 i 2 d, cover the entire set S and, thus, it covers the ball B. Clearly, each such ball has radius at most ρ/2. Thus, also in this case, B can be covered by at most 2 d balls of radius ρ/2. Theorem 4 Let λ 2 be an integer, let d = log λ, let k be a positive integer, and let n = 2 kd. There exists a metric space S of size n and doubling dimension d, such that the following are true: 1. Both the radius and diameter of S are equal to R = 2 k. ( 2. DP(S) = ) /2 d n 2. In particular, if d = 1, then DP(S) = n 2 /4. 3. wt(ann (S)) = 2Rn 1 1/d. In particular, if d = 1, then wt(ann (S)) = 2R. 4. If d = 1, then wt(mst (S)) = R log n. 5. If d > 1, then wt(mst (S)) 2Rn 1 1/d. Proof. Let S be the metric space of size m = 2 d, in which the distance between any two distinct points is equal to 2. The doubling dimension of S is equal to d, and both its radius and diameter are equal to R = 2. By applying the above construction k 1 times, we obtain a metric space S of size n = 2 kd, doubling dimension d, radius R = 2 k and diameter R = 2 k, proving the first claim. The second claim follows from (4). Since wt(ann (S )) = 2m = 2 2 d, it follows from (5) that wt(ann (S)) = 2 2 kd = 2n. Since n = 2 kd and R = 2 k, we have Rn 1 1/d = n. Therefore, we have wt(ann (S)) = 2Rn 1 1/d, proving the third claim. The fourth claim follows from the recurrence in (6). Finally, the fifth claim follows from the third one, because MST (S) contains ANN (S). 4 Concluding remarks We have shown that metric spaces of low doubling dimension are very different than low-dimensional Euclidean spaces. In this final section, we present 12

13 some more consequences of Lemma 1 for metric spaces of doubling dimension 1. Lemma 3 Let S be a finite metric space of doubling dimension 1, and let p, q, and r be three points in S such that (p, q) is an edge in MST (S) and pq pr qr. Then pr 2 pq. Proof. Let l = pq, l = pr, and l = qr. Thus, we have l l l. Let B be the ball with center p and radius l. This ball contains the points p, q, and r. Cover B by balls B 1 and B 2 centered at c 1 and c 2, respectively, and having radius l /2. We may assume without loss of generality that p is contained in B 1. Assume that r is in B 1. Then l = pr pc 1 + c 1 r l /2 + l /2 = l and, therefore, c 1 p = c 1 r = l /2 and pr = l, contradicting Lemma 1. Thus, r is in B 2. If we assume that q is in B 2, then, by a similar argument, it follows that c 2 q = c 2 r = l /2 and qr = l, contradicting Lemma 1. Thus, q is in B 1. We have shown that both p and q are in B 1, and that r is in B 2. Let T = MST (S). By removing the edge (p, q) from T, we obtain two trees T 1 and T 2, where p is a node in T 1 and q is a node in T 2. We may assume without loss of generality that c 1 is a node in T 1. Observe that (c 1, q) is not an edge in T. Let T be the tree obtained from T by replacing the edge (p, q) by the edge (c 1, q). Since T is a minimum spanning tree of S, it follows that l = pq c 1 q. On the other hand, since q is in B 1, we have c 1 q l /2. It follows that l 2l. We now use Lemma 3 to prove the following claims. First, by following any (directed) path in ANN (S), edges decrease in weight by a factor of at least 2. Second, for any point in S, any two incoming edges in ANN (S) differ in weight by a factor of at least 2. Finally, any two edges in MST (S) that share a point differ in length by a factor of at least 2. Lemma 4 Let S be a finite metric space of doubling dimension 1, and let p, q, and r be three pairwise distinct points in S. 1. If q = NN S (p) and p = NN S (r), then pr 2 pq. 13

14 2. If p = NN S (q), p = NN S (r), and pq pr, then pr 2 pq. 3. If both (p, q) and (p, r) are edges in MST (S) and pq pr, then pr 2 pq. Proof. The assumptions in the first and second claims imply that (p, q) is an edge in MST (S) and pq pr qr. Therefore, these two claims follow from Lemma 3. For the third claim, observe that (q, r) is not an edge in MST (S). It follows that pq pr qr and, therefore, the claim also follows from Lemma 3. References [1] P. Assouad. Plongements lipschitziens dans R N. Bulletin de la Société Mathématique de France, 111: , [2] L. Few. The shortest path and the shortest road through n points in a region. Mathematika, 2: , [3] S. Har-Peled and M. Mendel. Fast construction of nets in low-dimensional metrics and their applications. SIAM Journal on Computing, 35: , [4] J. Heinonen. Lectures on Analysis on Metric Spaces. Springer-Verlag, Berlin, [5] J. Pach and P. K. Agarwal. Combinatorial Geometry. John Wiley & Sons, New York, [6] M. Smid. The weak gap property in metric spaces of bounded doubling dimension. In Efficient Algorithms, Essays Dedicated to Kurt Mehlhorn on the Occasion of His 60th Birthday, volume 5760 of Lecture Notes in Computer Science, pages , Berlin, Springer-Verlag. [7] W. D. Smith. Studies in Computational Geometry Motivated by Mesh Generation. Ph.D. thesis, Priceton University, Princeton, NJ,

15 [8] J. M. Steele and T. L. Snyder. Worst-case growth rates of some classical problems of combinatorial optimization. SIAM Journal on Computing, 18: ,

COMBINATORIAL PROPERTIES OF THE HIGMAN-SIMS GRAPH. 1. Introduction

COMBINATORIAL PROPERTIES OF THE HIGMAN-SIMS GRAPH. 1. Introduction COMBINATORIAL PROPERTIES OF THE HIGMAN-SIMS GRAPH ZACHARY ABEL 1. Introduction In this survey we discuss properties of the Higman-Sims graph, which has 100 vertices, 1100 edges, and is 22 regular. In fact

More information

All trees contain a large induced subgraph having all degrees 1 (mod k)

All trees contain a large induced subgraph having all degrees 1 (mod k) All trees contain a large induced subgraph having all degrees 1 (mod k) David M. Berman, A.J. Radcliffe, A.D. Scott, Hong Wang, and Larry Wargo *Department of Mathematics University of New Orleans New

More information

Well-Separated Pair Decomposition for the Unit-disk Graph Metric and its Applications

Well-Separated Pair Decomposition for the Unit-disk Graph Metric and its Applications Well-Separated Pair Decomposition for the Unit-disk Graph Metric and its Applications Jie Gao Department of Computer Science Stanford University Joint work with Li Zhang Systems Research Center Hewlett-Packard

More information

The positive minimum degree game on sparse graphs

The positive minimum degree game on sparse graphs The positive minimum degree game on sparse graphs József Balogh Department of Mathematical Sciences University of Illinois, USA jobal@math.uiuc.edu András Pluhár Department of Computer Science University

More information

THE BANACH CONTRACTION PRINCIPLE. Contents

THE BANACH CONTRACTION PRINCIPLE. Contents THE BANACH CONTRACTION PRINCIPLE ALEX PONIECKI Abstract. This paper will study contractions of metric spaces. To do this, we will mainly use tools from topology. We will give some examples of contractions,

More information

CS 598CSC: Combinatorial Optimization Lecture date: 2/4/2010

CS 598CSC: Combinatorial Optimization Lecture date: 2/4/2010 CS 598CSC: Combinatorial Optimization Lecture date: /4/010 Instructor: Chandra Chekuri Scribe: David Morrison Gomory-Hu Trees (The work in this section closely follows [3]) Let G = (V, E) be an undirected

More information

Mathematical induction. Niloufar Shafiei

Mathematical induction. Niloufar Shafiei Mathematical induction Niloufar Shafiei Mathematical induction Mathematical induction is an extremely important proof technique. Mathematical induction can be used to prove results about complexity of

More information

Follow links for Class Use and other Permissions. For more information send email to: permissions@pupress.princeton.edu

Follow links for Class Use and other Permissions. For more information send email to: permissions@pupress.princeton.edu COPYRIGHT NOTICE: Ariel Rubinstein: Lecture Notes in Microeconomic Theory is published by Princeton University Press and copyrighted, c 2006, by Princeton University Press. All rights reserved. No part

More information

Chapter 11. 11.1 Load Balancing. Approximation Algorithms. Load Balancing. Load Balancing on 2 Machines. Load Balancing: Greedy Scheduling

Chapter 11. 11.1 Load Balancing. Approximation Algorithms. Load Balancing. Load Balancing on 2 Machines. Load Balancing: Greedy Scheduling Approximation Algorithms Chapter Approximation Algorithms Q. Suppose I need to solve an NP-hard problem. What should I do? A. Theory says you're unlikely to find a poly-time algorithm. Must sacrifice one

More information

Every tree contains a large induced subgraph with all degrees odd

Every tree contains a large induced subgraph with all degrees odd Every tree contains a large induced subgraph with all degrees odd A.J. Radcliffe Carnegie Mellon University, Pittsburgh, PA A.D. Scott Department of Pure Mathematics and Mathematical Statistics University

More information

Math212a1010 Lebesgue measure.

Math212a1010 Lebesgue measure. Math212a1010 Lebesgue measure. October 19, 2010 Today s lecture will be devoted to Lebesgue measure, a creation of Henri Lebesgue, in his thesis, one of the most famous theses in the history of mathematics.

More information

Introduction to Graph Theory

Introduction to Graph Theory Introduction to Graph Theory Allen Dickson October 2006 1 The Königsberg Bridge Problem The city of Königsberg was located on the Pregel river in Prussia. The river divided the city into four separate

More information

A Turán Type Problem Concerning the Powers of the Degrees of a Graph

A Turán Type Problem Concerning the Powers of the Degrees of a Graph A Turán Type Problem Concerning the Powers of the Degrees of a Graph Yair Caro and Raphael Yuster Department of Mathematics University of Haifa-ORANIM, Tivon 36006, Israel. AMS Subject Classification:

More information

Triangle deletion. Ernie Croot. February 3, 2010

Triangle deletion. Ernie Croot. February 3, 2010 Triangle deletion Ernie Croot February 3, 2010 1 Introduction The purpose of this note is to give an intuitive outline of the triangle deletion theorem of Ruzsa and Szemerédi, which says that if G = (V,

More information

ON THE COMPLEXITY OF THE GAME OF SET. {kamalika,pbg,dratajcz,hoeteck}@cs.berkeley.edu

ON THE COMPLEXITY OF THE GAME OF SET. {kamalika,pbg,dratajcz,hoeteck}@cs.berkeley.edu ON THE COMPLEXITY OF THE GAME OF SET KAMALIKA CHAUDHURI, BRIGHTEN GODFREY, DAVID RATAJCZAK, AND HOETECK WEE {kamalika,pbg,dratajcz,hoeteck}@cs.berkeley.edu ABSTRACT. Set R is a card game played with a

More information

Shortcut sets for plane Euclidean networks (Extended abstract) 1

Shortcut sets for plane Euclidean networks (Extended abstract) 1 Shortcut sets for plane Euclidean networks (Extended abstract) 1 J. Cáceres a D. Garijo b A. González b A. Márquez b M. L. Puertas a P. Ribeiro c a Departamento de Matemáticas, Universidad de Almería,

More information

Product irregularity strength of certain graphs

Product irregularity strength of certain graphs Also available at http://amc.imfm.si ISSN 1855-3966 (printed edn.), ISSN 1855-3974 (electronic edn.) ARS MATHEMATICA CONTEMPORANEA 7 (014) 3 9 Product irregularity strength of certain graphs Marcin Anholcer

More information

INCIDENCE-BETWEENNESS GEOMETRY

INCIDENCE-BETWEENNESS GEOMETRY INCIDENCE-BETWEENNESS GEOMETRY MATH 410, CSUSM. SPRING 2008. PROFESSOR AITKEN This document covers the geometry that can be developed with just the axioms related to incidence and betweenness. The full

More information

x 2 f 2 e 1 e 4 e 3 e 2 q f 4 x 4 f 3 x 3

x 2 f 2 e 1 e 4 e 3 e 2 q f 4 x 4 f 3 x 3 Karl-Franzens-Universitat Graz & Technische Universitat Graz SPEZIALFORSCHUNGSBEREICH F 003 OPTIMIERUNG und KONTROLLE Projektbereich DISKRETE OPTIMIERUNG O. Aichholzer F. Aurenhammer R. Hainz New Results

More information

Graph Theory Problems and Solutions

Graph Theory Problems and Solutions raph Theory Problems and Solutions Tom Davis tomrdavis@earthlink.net http://www.geometer.org/mathcircles November, 005 Problems. Prove that the sum of the degrees of the vertices of any finite graph is

More information

Handout #Ch7 San Skulrattanakulchai Gustavus Adolphus College Dec 6, 2010. Chapter 7: Digraphs

Handout #Ch7 San Skulrattanakulchai Gustavus Adolphus College Dec 6, 2010. Chapter 7: Digraphs MCS-236: Graph Theory Handout #Ch7 San Skulrattanakulchai Gustavus Adolphus College Dec 6, 2010 Chapter 7: Digraphs Strong Digraphs Definitions. A digraph is an ordered pair (V, E), where V is the set

More information

Mathematical Induction

Mathematical Induction Mathematical Induction (Handout March 8, 01) The Principle of Mathematical Induction provides a means to prove infinitely many statements all at once The principle is logical rather than strictly mathematical,

More information

! Solve problem to optimality. ! Solve problem in poly-time. ! Solve arbitrary instances of the problem. !-approximation algorithm.

! Solve problem to optimality. ! Solve problem in poly-time. ! Solve arbitrary instances of the problem. !-approximation algorithm. Approximation Algorithms Chapter Approximation Algorithms Q Suppose I need to solve an NP-hard problem What should I do? A Theory says you're unlikely to find a poly-time algorithm Must sacrifice one of

More information

Ph.D. Thesis. Judit Nagy-György. Supervisor: Péter Hajnal Associate Professor

Ph.D. Thesis. Judit Nagy-György. Supervisor: Péter Hajnal Associate Professor Online algorithms for combinatorial problems Ph.D. Thesis by Judit Nagy-György Supervisor: Péter Hajnal Associate Professor Doctoral School in Mathematics and Computer Science University of Szeged Bolyai

More information

PROBLEM SET 7: PIGEON HOLE PRINCIPLE

PROBLEM SET 7: PIGEON HOLE PRINCIPLE PROBLEM SET 7: PIGEON HOLE PRINCIPLE The pigeonhole principle is the following observation: Theorem. Suppose that > kn marbles are distributed over n jars, then one jar will contain at least k + marbles.

More information

Lecture 15 An Arithmetic Circuit Lowerbound and Flows in Graphs

Lecture 15 An Arithmetic Circuit Lowerbound and Flows in Graphs CSE599s: Extremal Combinatorics November 21, 2011 Lecture 15 An Arithmetic Circuit Lowerbound and Flows in Graphs Lecturer: Anup Rao 1 An Arithmetic Circuit Lower Bound An arithmetic circuit is just like

More information

High degree graphs contain large-star factors

High degree graphs contain large-star factors High degree graphs contain large-star factors Dedicated to László Lovász, for his 60th birthday Noga Alon Nicholas Wormald Abstract We show that any finite simple graph with minimum degree d contains a

More information

Large induced subgraphs with all degrees odd

Large induced subgraphs with all degrees odd Large induced subgraphs with all degrees odd A.D. Scott Department of Pure Mathematics and Mathematical Statistics, University of Cambridge, England Abstract: We prove that every connected graph of order

More information

SCORE SETS IN ORIENTED GRAPHS

SCORE SETS IN ORIENTED GRAPHS Applicable Analysis and Discrete Mathematics, 2 (2008), 107 113. Available electronically at http://pefmath.etf.bg.ac.yu SCORE SETS IN ORIENTED GRAPHS S. Pirzada, T. A. Naikoo The score of a vertex v in

More information

Divide-and-Conquer. Three main steps : 1. divide; 2. conquer; 3. merge.

Divide-and-Conquer. Three main steps : 1. divide; 2. conquer; 3. merge. Divide-and-Conquer Three main steps : 1. divide; 2. conquer; 3. merge. 1 Let I denote the (sub)problem instance and S be its solution. The divide-and-conquer strategy can be described as follows. Procedure

More information

Cycles in a Graph Whose Lengths Differ by One or Two

Cycles in a Graph Whose Lengths Differ by One or Two Cycles in a Graph Whose Lengths Differ by One or Two J. A. Bondy 1 and A. Vince 2 1 LABORATOIRE DE MATHÉMATIQUES DISCRÉTES UNIVERSITÉ CLAUDE-BERNARD LYON 1 69622 VILLEURBANNE, FRANCE 2 DEPARTMENT OF MATHEMATICS

More information

arxiv:1203.1525v1 [math.co] 7 Mar 2012

arxiv:1203.1525v1 [math.co] 7 Mar 2012 Constructing subset partition graphs with strong adjacency and end-point count properties Nicolai Hähnle haehnle@math.tu-berlin.de arxiv:1203.1525v1 [math.co] 7 Mar 2012 March 8, 2012 Abstract Kim defined

More information

INDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS

INDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS INDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS STEVEN P. LALLEY AND ANDREW NOBEL Abstract. It is shown that there are no consistent decision rules for the hypothesis testing problem

More information

A Non-Linear Schema Theorem for Genetic Algorithms

A Non-Linear Schema Theorem for Genetic Algorithms A Non-Linear Schema Theorem for Genetic Algorithms William A Greene Computer Science Department University of New Orleans New Orleans, LA 70148 bill@csunoedu 504-280-6755 Abstract We generalize Holland

More information

CONTRIBUTIONS TO ZERO SUM PROBLEMS

CONTRIBUTIONS TO ZERO SUM PROBLEMS CONTRIBUTIONS TO ZERO SUM PROBLEMS S. D. ADHIKARI, Y. G. CHEN, J. B. FRIEDLANDER, S. V. KONYAGIN AND F. PAPPALARDI Abstract. A prototype of zero sum theorems, the well known theorem of Erdős, Ginzburg

More information

Homework until Test #2

Homework until Test #2 MATH31: Number Theory Homework until Test # Philipp BRAUN Section 3.1 page 43, 1. It has been conjectured that there are infinitely many primes of the form n. Exhibit five such primes. Solution. Five such

More information

PART I. THE REAL NUMBERS

PART I. THE REAL NUMBERS PART I. THE REAL NUMBERS This material assumes that you are already familiar with the real number system and the representation of the real numbers as points on the real line. I.1. THE NATURAL NUMBERS

More information

Collinear Points in Permutations

Collinear Points in Permutations Collinear Points in Permutations Joshua N. Cooper Courant Institute of Mathematics New York University, New York, NY József Solymosi Department of Mathematics University of British Columbia, Vancouver,

More information

SOLUTIONS TO ASSIGNMENT 1 MATH 576

SOLUTIONS TO ASSIGNMENT 1 MATH 576 SOLUTIONS TO ASSIGNMENT 1 MATH 576 SOLUTIONS BY OLIVIER MARTIN 13 #5. Let T be the topology generated by A on X. We want to show T = J B J where B is the set of all topologies J on X with A J. This amounts

More information

Vector Spaces II: Finite Dimensional Linear Algebra 1

Vector Spaces II: Finite Dimensional Linear Algebra 1 John Nachbar September 2, 2014 Vector Spaces II: Finite Dimensional Linear Algebra 1 1 Definitions and Basic Theorems. For basic properties and notation for R N, see the notes Vector Spaces I. Definition

More information

1. Prove that the empty set is a subset of every set.

1. Prove that the empty set is a subset of every set. 1. Prove that the empty set is a subset of every set. Basic Topology Written by Men-Gen Tsai email: b89902089@ntu.edu.tw Proof: For any element x of the empty set, x is also an element of every set since

More information

1. Determine all real numbers a, b, c, d that satisfy the following system of equations.

1. Determine all real numbers a, b, c, d that satisfy the following system of equations. altic Way 1999 Reykjavík, November 6, 1999 Problems 1. etermine all real numbers a, b, c, d that satisfy the following system of equations. abc + ab + bc + ca + a + b + c = 1 bcd + bc + cd + db + b + c

More information

princeton univ. F 13 cos 521: Advanced Algorithm Design Lecture 6: Provable Approximation via Linear Programming Lecturer: Sanjeev Arora

princeton univ. F 13 cos 521: Advanced Algorithm Design Lecture 6: Provable Approximation via Linear Programming Lecturer: Sanjeev Arora princeton univ. F 13 cos 521: Advanced Algorithm Design Lecture 6: Provable Approximation via Linear Programming Lecturer: Sanjeev Arora Scribe: One of the running themes in this course is the notion of

More information

The Union-Find Problem Kruskal s algorithm for finding an MST presented us with a problem in data-structure design. As we looked at each edge,

The Union-Find Problem Kruskal s algorithm for finding an MST presented us with a problem in data-structure design. As we looked at each edge, The Union-Find Problem Kruskal s algorithm for finding an MST presented us with a problem in data-structure design. As we looked at each edge, cheapest first, we had to determine whether its two endpoints

More information

Practical Guide to the Simplex Method of Linear Programming

Practical Guide to the Simplex Method of Linear Programming Practical Guide to the Simplex Method of Linear Programming Marcel Oliver Revised: April, 0 The basic steps of the simplex algorithm Step : Write the linear programming problem in standard form Linear

More information

Open and Closed Sets

Open and Closed Sets Open and Closed Sets Definition: A subset S of a metric space (X, d) is open if it contains an open ball about each of its points i.e., if x S : ɛ > 0 : B(x, ɛ) S. (1) Theorem: (O1) and X are open sets.

More information

Induction. Margaret M. Fleck. 10 October These notes cover mathematical induction and recursive definition

Induction. Margaret M. Fleck. 10 October These notes cover mathematical induction and recursive definition Induction Margaret M. Fleck 10 October 011 These notes cover mathematical induction and recursive definition 1 Introduction to induction At the start of the term, we saw the following formula for computing

More information

Sample Induction Proofs

Sample Induction Proofs Math 3 Worksheet: Induction Proofs III, Sample Proofs A.J. Hildebrand Sample Induction Proofs Below are model solutions to some of the practice problems on the induction worksheets. The solutions given

More information

SHARP BOUNDS FOR THE SUM OF THE SQUARES OF THE DEGREES OF A GRAPH

SHARP BOUNDS FOR THE SUM OF THE SQUARES OF THE DEGREES OF A GRAPH 31 Kragujevac J. Math. 25 (2003) 31 49. SHARP BOUNDS FOR THE SUM OF THE SQUARES OF THE DEGREES OF A GRAPH Kinkar Ch. Das Department of Mathematics, Indian Institute of Technology, Kharagpur 721302, W.B.,

More information

A Note on Maximum Independent Sets in Rectangle Intersection Graphs

A Note on Maximum Independent Sets in Rectangle Intersection Graphs A Note on Maximum Independent Sets in Rectangle Intersection Graphs Timothy M. Chan School of Computer Science University of Waterloo Waterloo, Ontario N2L 3G1, Canada tmchan@uwaterloo.ca September 12,

More information

A Sublinear Bipartiteness Tester for Bounded Degree Graphs

A Sublinear Bipartiteness Tester for Bounded Degree Graphs A Sublinear Bipartiteness Tester for Bounded Degree Graphs Oded Goldreich Dana Ron February 5, 1998 Abstract We present a sublinear-time algorithm for testing whether a bounded degree graph is bipartite

More information

No: 10 04. Bilkent University. Monotonic Extension. Farhad Husseinov. Discussion Papers. Department of Economics

No: 10 04. Bilkent University. Monotonic Extension. Farhad Husseinov. Discussion Papers. Department of Economics No: 10 04 Bilkent University Monotonic Extension Farhad Husseinov Discussion Papers Department of Economics The Discussion Papers of the Department of Economics are intended to make the initial results

More information

Computer Science Department. Technion - IIT, Haifa, Israel. Itai and Rodeh [IR] have proved that for any 2-connected graph G and any vertex s G there

Computer Science Department. Technion - IIT, Haifa, Israel. Itai and Rodeh [IR] have proved that for any 2-connected graph G and any vertex s G there - 1 - THREE TREE-PATHS Avram Zehavi Alon Itai Computer Science Department Technion - IIT, Haifa, Israel Abstract Itai and Rodeh [IR] have proved that for any 2-connected graph G and any vertex s G there

More information

Cycles and clique-minors in expanders

Cycles and clique-minors in expanders Cycles and clique-minors in expanders Benny Sudakov UCLA and Princeton University Expanders Definition: The vertex boundary of a subset X of a graph G: X = { all vertices in G\X with at least one neighbor

More information

Analysis of Algorithms, I

Analysis of Algorithms, I Analysis of Algorithms, I CSOR W4231.002 Eleni Drinea Computer Science Department Columbia University Thursday, February 26, 2015 Outline 1 Recap 2 Representing graphs 3 Breadth-first search (BFS) 4 Applications

More information

Tilings of the sphere with right triangles III: the asymptotically obtuse families

Tilings of the sphere with right triangles III: the asymptotically obtuse families Tilings of the sphere with right triangles III: the asymptotically obtuse families Robert J. MacG. Dawson Department of Mathematics and Computing Science Saint Mary s University Halifax, Nova Scotia, Canada

More information

Clique coloring B 1 -EPG graphs

Clique coloring B 1 -EPG graphs Clique coloring B 1 -EPG graphs Flavia Bonomo a,c, María Pía Mazzoleni b,c, and Maya Stein d a Departamento de Computación, FCEN-UBA, Buenos Aires, Argentina. b Departamento de Matemática, FCE-UNLP, La

More information

LEARNING OBJECTIVES FOR THIS CHAPTER

LEARNING OBJECTIVES FOR THIS CHAPTER CHAPTER 2 American mathematician Paul Halmos (1916 2006), who in 1942 published the first modern linear algebra book. The title of Halmos s book was the same as the title of this chapter. Finite-Dimensional

More information

Baltic Way 1995. Västerås (Sweden), November 12, 1995. Problems and solutions

Baltic Way 1995. Västerås (Sweden), November 12, 1995. Problems and solutions Baltic Way 995 Västerås (Sweden), November, 995 Problems and solutions. Find all triples (x, y, z) of positive integers satisfying the system of equations { x = (y + z) x 6 = y 6 + z 6 + 3(y + z ). Solution.

More information

Best Monotone Degree Bounds for Various Graph Parameters

Best Monotone Degree Bounds for Various Graph Parameters Best Monotone Degree Bounds for Various Graph Parameters D. Bauer Department of Mathematical Sciences Stevens Institute of Technology Hoboken, NJ 07030 S. L. Hakimi Department of Electrical and Computer

More information

GRAPH THEORY LECTURE 4: TREES

GRAPH THEORY LECTURE 4: TREES GRAPH THEORY LECTURE 4: TREES Abstract. 3.1 presents some standard characterizations and properties of trees. 3.2 presents several different types of trees. 3.7 develops a counting method based on a bijection

More information

Counting Falsifying Assignments of a 2-CF via Recurrence Equations

Counting Falsifying Assignments of a 2-CF via Recurrence Equations Counting Falsifying Assignments of a 2-CF via Recurrence Equations Guillermo De Ita Luna, J. Raymundo Marcial Romero, Fernando Zacarias Flores, Meliza Contreras González and Pedro Bello López Abstract

More information

Induction Problems. Tom Davis November 7, 2005

Induction Problems. Tom Davis  November 7, 2005 Induction Problems Tom Davis tomrdavis@earthlin.net http://www.geometer.org/mathcircles November 7, 2005 All of the following problems should be proved by mathematical induction. The problems are not necessarily

More information

CS 103X: Discrete Structures Homework Assignment 3 Solutions

CS 103X: Discrete Structures Homework Assignment 3 Solutions CS 103X: Discrete Structures Homework Assignment 3 s Exercise 1 (20 points). On well-ordering and induction: (a) Prove the induction principle from the well-ordering principle. (b) Prove the well-ordering

More information

The Prime Numbers. Definition. A prime number is a positive integer with exactly two positive divisors.

The Prime Numbers. Definition. A prime number is a positive integer with exactly two positive divisors. The Prime Numbers Before starting our study of primes, we record the following important lemma. Recall that integers a, b are said to be relatively prime if gcd(a, b) = 1. Lemma (Euclid s Lemma). If gcd(a,

More information

Exponential time algorithms for graph coloring

Exponential time algorithms for graph coloring Exponential time algorithms for graph coloring Uriel Feige Lecture notes, March 14, 2011 1 Introduction Let [n] denote the set {1,..., k}. A k-labeling of vertices of a graph G(V, E) is a function V [k].

More information

Connectivity and cuts

Connectivity and cuts Math 104, Graph Theory February 19, 2013 Measure of connectivity How connected are each of these graphs? > increasing connectivity > I G 1 is a tree, so it is a connected graph w/minimum # of edges. Every

More information

Solutions for Practice problems on proofs

Solutions for Practice problems on proofs Solutions for Practice problems on proofs Definition: (even) An integer n Z is even if and only if n = 2m for some number m Z. Definition: (odd) An integer n Z is odd if and only if n = 2m + 1 for some

More information

Applied Algorithm Design Lecture 5

Applied Algorithm Design Lecture 5 Applied Algorithm Design Lecture 5 Pietro Michiardi Eurecom Pietro Michiardi (Eurecom) Applied Algorithm Design Lecture 5 1 / 86 Approximation Algorithms Pietro Michiardi (Eurecom) Applied Algorithm Design

More information

CSC2420 Fall 2012: Algorithm Design, Analysis and Theory

CSC2420 Fall 2012: Algorithm Design, Analysis and Theory CSC2420 Fall 2012: Algorithm Design, Analysis and Theory Allan Borodin November 15, 2012; Lecture 10 1 / 27 Randomized online bipartite matching and the adwords problem. We briefly return to online algorithms

More information

Uniform Multicommodity Flow through the Complete Graph with Random Edge-capacities

Uniform Multicommodity Flow through the Complete Graph with Random Edge-capacities Combinatorics, Probability and Computing (2005) 00, 000 000. c 2005 Cambridge University Press DOI: 10.1017/S0000000000000000 Printed in the United Kingdom Uniform Multicommodity Flow through the Complete

More information

The degree, size and chromatic index of a uniform hypergraph

The degree, size and chromatic index of a uniform hypergraph The degree, size and chromatic index of a uniform hypergraph Noga Alon Jeong Han Kim Abstract Let H be a k-uniform hypergraph in which no two edges share more than t common vertices, and let D denote the

More information

Analysis of Server Provisioning for Distributed Interactive Applications

Analysis of Server Provisioning for Distributed Interactive Applications Analysis of Server Provisioning for Distributed Interactive Applications Hanying Zheng and Xueyan Tang Abstract Increasing geographical spreads of modern distributed interactive applications DIAs make

More information

Odd induced subgraphs in graphs of maximum degree three

Odd induced subgraphs in graphs of maximum degree three Odd induced subgraphs in graphs of maximum degree three David M. Berman, Hong Wang, and Larry Wargo Department of Mathematics University of New Orleans New Orleans, Louisiana, USA 70148 Abstract A long-standing

More information

2.3 Convex Constrained Optimization Problems

2.3 Convex Constrained Optimization Problems 42 CHAPTER 2. FUNDAMENTAL CONCEPTS IN CONVEX OPTIMIZATION Theorem 15 Let f : R n R and h : R R. Consider g(x) = h(f(x)) for all x R n. The function g is convex if either of the following two conditions

More information

On the k-path cover problem for cacti

On the k-path cover problem for cacti On the k-path cover problem for cacti Zemin Jin and Xueliang Li Center for Combinatorics and LPMC Nankai University Tianjin 300071, P.R. China zeminjin@eyou.com, x.li@eyou.com Abstract In this paper we

More information

MA651 Topology. Lecture 6. Separation Axioms.

MA651 Topology. Lecture 6. Separation Axioms. MA651 Topology. Lecture 6. Separation Axioms. This text is based on the following books: Fundamental concepts of topology by Peter O Neil Elements of Mathematics: General Topology by Nicolas Bourbaki Counterexamples

More information

Basic Concepts of Point Set Topology Notes for OU course Math 4853 Spring 2011

Basic Concepts of Point Set Topology Notes for OU course Math 4853 Spring 2011 Basic Concepts of Point Set Topology Notes for OU course Math 4853 Spring 2011 A. Miller 1. Introduction. The definitions of metric space and topological space were developed in the early 1900 s, largely

More information

! Solve problem to optimality. ! Solve problem in poly-time. ! Solve arbitrary instances of the problem. #-approximation algorithm.

! Solve problem to optimality. ! Solve problem in poly-time. ! Solve arbitrary instances of the problem. #-approximation algorithm. Approximation Algorithms 11 Approximation Algorithms Q Suppose I need to solve an NP-hard problem What should I do? A Theory says you're unlikely to find a poly-time algorithm Must sacrifice one of three

More information

Near Optimal Solutions

Near Optimal Solutions Near Optimal Solutions Many important optimization problems are lacking efficient solutions. NP-Complete problems unlikely to have polynomial time solutions. Good heuristics important for such problems.

More information

Graphs without proper subgraphs of minimum degree 3 and short cycles

Graphs without proper subgraphs of minimum degree 3 and short cycles Graphs without proper subgraphs of minimum degree 3 and short cycles Lothar Narins, Alexey Pokrovskiy, Tibor Szabó Department of Mathematics, Freie Universität, Berlin, Germany. August 22, 2014 Abstract

More information

arxiv: v2 [math.co] 30 Nov 2015

arxiv: v2 [math.co] 30 Nov 2015 PLANAR GRAPH IS ON FIRE PRZEMYSŁAW GORDINOWICZ arxiv:1311.1158v [math.co] 30 Nov 015 Abstract. Let G be any connected graph on n vertices, n. Let k be any positive integer. Suppose that a fire breaks out

More information

Metric Spaces Joseph Muscat 2003 (Last revised May 2009)

Metric Spaces Joseph Muscat 2003 (Last revised May 2009) 1 Distance J Muscat 1 Metric Spaces Joseph Muscat 2003 (Last revised May 2009) (A revised and expanded version of these notes are now published by Springer.) 1 Distance A metric space can be thought of

More information

SEQUENCES OF MAXIMAL DEGREE VERTICES IN GRAPHS. Nickolay Khadzhiivanov, Nedyalko Nenov

SEQUENCES OF MAXIMAL DEGREE VERTICES IN GRAPHS. Nickolay Khadzhiivanov, Nedyalko Nenov Serdica Math. J. 30 (2004), 95 102 SEQUENCES OF MAXIMAL DEGREE VERTICES IN GRAPHS Nickolay Khadzhiivanov, Nedyalko Nenov Communicated by V. Drensky Abstract. Let Γ(M) where M V (G) be the set of all vertices

More information

k, then n = p2α 1 1 pα k

k, then n = p2α 1 1 pα k Powers of Integers An integer n is a perfect square if n = m for some integer m. Taking into account the prime factorization, if m = p α 1 1 pα k k, then n = pα 1 1 p α k k. That is, n is a perfect square

More information

Lecture 3: Finding integer solutions to systems of linear equations

Lecture 3: Finding integer solutions to systems of linear equations Lecture 3: Finding integer solutions to systems of linear equations Algorithmic Number Theory (Fall 2014) Rutgers University Swastik Kopparty Scribe: Abhishek Bhrushundi 1 Overview The goal of this lecture

More information

Continued Fractions and the Euclidean Algorithm

Continued Fractions and the Euclidean Algorithm Continued Fractions and the Euclidean Algorithm Lecture notes prepared for MATH 326, Spring 997 Department of Mathematics and Statistics University at Albany William F Hammond Table of Contents Introduction

More information

ZERO-DIVISOR GRAPHS OF POLYNOMIALS AND POWER SERIES OVER COMMUTATIVE RINGS

ZERO-DIVISOR GRAPHS OF POLYNOMIALS AND POWER SERIES OVER COMMUTATIVE RINGS ZERO-DIVISOR GRAPHS OF POLYNOMIALS AND POWER SERIES OVER COMMUTATIVE RINGS M. AXTELL, J. COYKENDALL, AND J. STICKLES Abstract. We recall several results of zero divisor graphs of commutative rings. We

More information

Approximated Distributed Minimum Vertex Cover Algorithms for Bounded Degree Graphs

Approximated Distributed Minimum Vertex Cover Algorithms for Bounded Degree Graphs Approximated Distributed Minimum Vertex Cover Algorithms for Bounded Degree Graphs Yong Zhang 1.2, Francis Y.L. Chin 2, and Hing-Fung Ting 2 1 College of Mathematics and Computer Science, Hebei University,

More information

Solving Geometric Problems with the Rotating Calipers *

Solving Geometric Problems with the Rotating Calipers * Solving Geometric Problems with the Rotating Calipers * Godfried Toussaint School of Computer Science McGill University Montreal, Quebec, Canada ABSTRACT Shamos [1] recently showed that the diameter of

More information

1 The Line vs Point Test

1 The Line vs Point Test 6.875 PCP and Hardness of Approximation MIT, Fall 2010 Lecture 5: Low Degree Testing Lecturer: Dana Moshkovitz Scribe: Gregory Minton and Dana Moshkovitz Having seen a probabilistic verifier for linearity

More information

ON INDUCED SUBGRAPHS WITH ALL DEGREES ODD. 1. Introduction

ON INDUCED SUBGRAPHS WITH ALL DEGREES ODD. 1. Introduction ON INDUCED SUBGRAPHS WITH ALL DEGREES ODD A.D. SCOTT Abstract. Gallai proved that the vertex set of any graph can be partitioned into two sets, each inducing a subgraph with all degrees even. We prove

More information

(Basic definitions and properties; Separation theorems; Characterizations) 1.1 Definition, examples, inner description, algebraic properties

(Basic definitions and properties; Separation theorems; Characterizations) 1.1 Definition, examples, inner description, algebraic properties Lecture 1 Convex Sets (Basic definitions and properties; Separation theorems; Characterizations) 1.1 Definition, examples, inner description, algebraic properties 1.1.1 A convex set In the school geometry

More information

POLYNOMIAL RINGS AND UNIQUE FACTORIZATION DOMAINS

POLYNOMIAL RINGS AND UNIQUE FACTORIZATION DOMAINS POLYNOMIAL RINGS AND UNIQUE FACTORIZATION DOMAINS RUSS WOODROOFE 1. Unique Factorization Domains Throughout the following, we think of R as sitting inside R[x] as the constant polynomials (of degree 0).

More information

Approximation Algorithms

Approximation Algorithms Approximation Algorithms or: How I Learned to Stop Worrying and Deal with NP-Completeness Ong Jit Sheng, Jonathan (A0073924B) March, 2012 Overview Key Results (I) General techniques: Greedy algorithms

More information

Graphical degree sequences and realizations

Graphical degree sequences and realizations swap Graphical and realizations Péter L. Erdös Alfréd Rényi Institute of Mathematics Hungarian Academy of Sciences MAPCON 12 MPIPKS - Dresden, May 15, 2012 swap Graphical and realizations Péter L. Erdös

More information

On Integer Additive Set-Indexers of Graphs

On Integer Additive Set-Indexers of Graphs On Integer Additive Set-Indexers of Graphs arxiv:1312.7672v4 [math.co] 2 Mar 2014 N K Sudev and K A Germina Abstract A set-indexer of a graph G is an injective set-valued function f : V (G) 2 X such that

More information

136 CHAPTER 4. INDUCTION, GRAPHS AND TREES

136 CHAPTER 4. INDUCTION, GRAPHS AND TREES 136 TER 4. INDUCTION, GRHS ND TREES 4.3 Graphs In this chapter we introduce a fundamental structural idea of discrete mathematics, that of a graph. Many situations in the applications of discrete mathematics

More information

arxiv:math/0202219v1 [math.co] 21 Feb 2002

arxiv:math/0202219v1 [math.co] 21 Feb 2002 RESTRICTED PERMUTATIONS BY PATTERNS OF TYPE (2, 1) arxiv:math/0202219v1 [math.co] 21 Feb 2002 TOUFIK MANSOUR LaBRI (UMR 5800), Université Bordeaux 1, 351 cours de la Libération, 33405 Talence Cedex, France

More information

The Goldberg Rao Algorithm for the Maximum Flow Problem

The Goldberg Rao Algorithm for the Maximum Flow Problem The Goldberg Rao Algorithm for the Maximum Flow Problem COS 528 class notes October 18, 2006 Scribe: Dávid Papp Main idea: use of the blocking flow paradigm to achieve essentially O(min{m 2/3, n 1/2 }

More information